Announcement

Collapse
No announcement yet.

Jolla Announces Their First Phone

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    If anyone in this thread has doubts that Sailfish OS will be opened:
    Our goal with the Sailfish OS is to develop an open source operating system in co-operation with the community, thus ensuring the development of a best-of-breed operating system.
    [...]
    from https://sailfishos.org/about-license.html

    So the claim of openness is not only an ambiguous twitter post.

    Comment


    • #32
      Originally posted by Armin View Post
      Don't forget it's a mobile phone not a super computer!
      People do 1080p video playback, and 1080p video recording.
      Maybe at same time that they're downloading and uploading stuff, perhaps streaming stuff, and maybe the OS is updating packages too.

      People also can do stuff like connect their mobile phone over Wi-Fi and be a host Access Point for other devices, or to connect to a TV over DLNA and stream video and audio to the TV.

      People also can connect a HDMI cable to the phone, and mouse and keyboard over Bluetooth, then use their smartphone as a computer.
      Then you might want to boot GNU/Linux and do some work in with GIMP.

      Comment


      • #33
        Mer+Nemo will give you an compatible open source solution.
        Even if Sailfish OS won't open source everything you will still be able to use an compatible open source solution.

        Originally posted by https://sailfishos.org/wiki/QA
        What is the relationship of Sailfish OS and Nemo Mobile project?

        We are using parts of Nemo Mobile in Sailfish OS. Specifically Nemo Mobile has a UI and a set of applications that we won't use in the final Jolla product nor in Sailfish OS. We are contributing to Nemo Mobile and our intention is to keep Sailfish OS compatible with Nemo Mobile and vice versa.


        What does Sailfish OS have on top of Mer?

        Mer provides a functional core Linux stack that boots up to the screen. Mer does not provide HW adaptation nor a functional UI. There is where Sailfish fills the gap and provide you with a productized mobile OS.
        The HW adaptation isn't completely up to Jolla (if a HW manufacture provide OSS drivers or not isn't up to Jolla) but they do develop libhybris that aims to make HW adaptation easier.
        Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

        For the UI there is an open replacement called Nemo.

        Originally posted by https://sailfishos.org/wiki/QA
        What is your open source contribution model?

        Sailfish OS is based on Mer and Qt which are already existing and known open-source projects with good contribution policies.

        Do you use in-house patched Qt version with Jolla specific modifications?

        We use the fully open version of Qt delivered by Mer; like many distributions this carries small patches to fix bugs, enhance performance or backport features. Mer tracks upstream Qt very closely.


        How isn't Sailfish OS with the option of using mer+nemo open enough?
        On my N9 i can dual boot Harmattan and mer+nemo, since Jolla aims to be more open then the Nokia N9 i find it hard to believe that i wouldn't be able to dual boot on the new device.

        Comment


        • #34
          Originally posted by uid313 View Post
          People do 1080p video playback with a dedicated chip, and 1080p video recording with a dedicated chip.
          Maybe at same time that they're downloading and uploading stuff using the 4G chip, perhaps streaming stuff using the 4G chip, and maybe the OS is updating packages too when the phone is idle.

          People also can do stuff like connect their mobile phone over Wi-Fi and be a host Access Point for other devices using mostly the Wifi chip, or to connect to a TV over DLNA and stream video and audio to the TV without transcoding or using hardware transcoding.

          People also can connect a HDMI cable to the phone, and mouse and keyboard over Bluetooth, then use their smartphone as a computer.
          Then you might want to boot GNU/Linux and do some work in with GIMP, which will be extra slow on an mobile arm chip with less than 2Bb of ram.
          fixed for you

          More seriously, a core for the OS and a core for the current app is good for responsiveness, but that's about all. The only thing more cores could be useful for would be games, possibly (but as it's much more difficult to ramp up AI than graphics, mobile games tend to be easy on CPU).
          For anything that is both computationally intensive and a real use case for mobile, a dedicated chip is always a better solution, especially given how weak the cores are in the first place.
          Case in point: if the media chip doesn't support one codec, you won't read a 1080p encoded video (with such codec), not with any smartphone on the market today.

          Comment


          • #35
            Originally posted by erendorn View Post
            The only thing more cores could be useful for would be games, possibly (but as it's much more difficult to ramp up AI than graphics, mobile games tend to be easy on CPU).
            I disagree. There are many use cases where more cores help. All the following can benefit from more cores:
            • Primary one is the browser, where you can have several tabs open. Background tabs download and render while you interact with the foreground one.
            • Then photo editing: Users apply filter/transformation to photos (panorama, Instagram etc).
            • Video transcode/editing with immediate upload to YouTube after capture.
            • Using the phone as desktop (e.g. Ubuntu for Android) by connecting a monitor via MHD, and keyboard/mouse.
            • Augmented reality with face/object recognition.

            Comment


            • #36
              Originally posted by chithanh View Post
              I disagree. There are many use cases where more cores help. All the following can benefit from more cores:
              • Primary one is the browser, where you can have several tabs open. Background tabs download and render while you interact with the foreground one.
              • Then photo editing: Users apply filter/transformation to photos (panorama, Instagram etc).
              • Video transcode/editing with immediate upload to YouTube after capture.
              • Using the phone as desktop (e.g. Ubuntu for Android) by connecting a monitor via MHD, and keyboard/mouse.
              • Augmented reality with face/object recognition.
              Same comments:
              • Not sure you need to render several (more than 1, less than 4) background tabs all at the same time so often, but lets keep this one.
              • if you want acceptable perf, you need dedicated chip, or openGL ES (if its parallelisable, you'll get x10 to x100 instead of x2)
              • dedicated chip (no CPU encode/decode on phone, whatever the number of cores). Transferring data is not CPU intensive.
              • each core is too slow anyway, adding more cores won't get your linear applications much faster. Still the low ram and not so fast local data constraints. You'll have a less-than-2-year-old-atom-netbook experience with even the highest end hardware.
              • not parallel

              Adding cores after the first 2 is most useless given the ram and the actual power of each core on a mobile phone.

              Comment


              • #37
                Originally posted by erendorn View Post
                Adding cores after the first 2 is most useless given the ram and the actual power of each core on a mobile phone.
                You can accelerate image manipulation with GPGPU or similar, but not all manipulations are equally suited for that.
                Video transcode is possible with dedicated hardware, but postprocessing is mostly not. I admit that it can be done in GPU shaders to some degree. Audio postprocessing on the GPU is at least not widespread yet I think
                You typically have more applications open on a desktop than on a mobile phone, so more cores (and indeed more RAM too) help.
                If you want to capture a video and automatically send the YouTube link to all of your friends which appear therein, then individual cores can start looking for recognizable faces at every I-frame.

                Comment


                • #38
                  +1

                  Phone is by its very nature ultra-low-power platform. It is nice to have adequate CPU power so that your camera app doesn't take a eternity to load and snap a picture, nut much more than that can quickly have negative final value. Everything in it is severely thermally, power-wise and energy-wise limited.

                  For great majority of users it doesn't pay to strive for "everything-and-even-kitchen-sink" approach. If you need power of laptop, use laptop.

                  If I could choose one change on the phone, I would strike out that micro-sdhc card and replace it with optional proprietary FLASH module with high bandwidth. I would use seentially just FLASH chips on wide parallel bus on board without any other interface chip, so that phone CPU could have direct wide full-bandwidth to the flash.

                  It would be more expensive than elcheapo sdhc, but not especially so. But bandwidth gain would probably be spectacular.

                  It would also be nice to have some fast system interface, like PCIex1, so that phone could be coupled to some eg. measurement equipment etc...

                  Comment


                  • #39
                    Originally posted by Brane215 View Post
                    If I could choose one change on the phone, I would strike out that micro-sdhc card and replace it with optional proprietary FLASH module with high bandwidth. I would use seentially just FLASH chips on wide parallel bus on board without any other interface chip, so that phone CPU could have direct wide full-bandwidth to the flash.

                    It would be more expensive than elcheapo sdhc, but not especially so. But bandwidth gain would probably be spectacular.

                    It would also be nice to have some fast system interface, like PCIex1, so that phone could be coupled to some eg. measurement equipment etc...
                    Who knows, maybe The Other Side includes something like that. Hard to tell at this point.

                    Comment


                    • #40
                      As for CPU power, I need enough to make browsing the net smooth. Anything beyond that is waste of money/not interested.

                      Comment

                      Working...
                      X